AgroFix

Here’s a polished version you can use for your public project page about your AI-based Agricultural Advisory project. I’ve formatted it in Markdown with sections ready for LaTeX if needed:

Project Story

Inspiration

The inspiration for this project came from witnessing the challenges faced by farmers in accessing timely and accurate agricultural advice. Many farmers rely on traditional methods and fragmented information, which can lead to poor crop yields, pest outbreaks, and inefficient use of resources. I wanted to leverage AI to provide a personalized, real-time advisory platform that could help farmers make informed decisions based on soil health, weather patterns, and crop disease detection.

What I Learned

During this project, I gained hands-on experience in:

• Building AI models for image recognition to detect crop diseases.
• Implementing RAG (Retrieval-Augmented Generation) techniques to provide context-aware recommendations.
• Orchestrating multiple AI modules and databases efficiently.
• Handling real-world agricultural data, which is often noisy and incomplete.
• Deploying AI models in a way that is scalable and accessible via web interfaces.

How I Built It

The platform consists of several integrated modules:

1. AI Orchestrator – Manages requests and decides which AI model or knowledge engine to use.
2. Context Memory Engine – Stores historical farmer queries and recommendations for personalized advice.
3. RAG Knowledge Engine – Uses vector embeddings and a document store to retrieve relevant agricultural information.
4. Vision Disease Detection Model – A ResNet-based model that analyzes crop images and identifies diseases.
5. Soil Health Analyzer – Processes soil sample data and provides nutrient recommendations.

The front-end is a web-based dashboard where farmers can input queries, upload crop images, and receive AI-powered recommendations instantly.

Challenges included handling incomplete soil and weather data, training the disease detection model with limited labeled images, and ensuring that the system can provide reliable advice under uncertain conditions.

Challenges

• Limited datasets for rare crop diseases.
• Integrating multiple AI models into a seamless workflow.
• Ensuring low-latency responses for farmers with limited internet connectivity.

--

Built With

• actions
• aws-lambda-(serverless-functions)-*-**apis-&-tools:**-openai-api-(llm-for-advisory)
• chroma-db-(vector-embeddings)-*-**cloud-services:**-aws-s3-(storage)
• ci/cd
• flask-*-**databases:**-postgresql-(context-memory)
• for
• github
• pytorch
• resnet-for-image-classification-*-**other:**-docker-for-containerization
• sentence-transformers-for-embeddings
• streamlit